Fatigue Crack Detection via Image Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting fatigue cracks in civil infrastructure, such as steel bridges, are often labor-intensive, prone to errors, and costly, especially when dealing with large-scale structures, as they require extensive human operation and complex monitoring systems, and struggle to distinguish true cracks from non-crack features like wires or corrosion marks.
Innovation Solution
A vision-based non-contact approach using image overlapping techniques that capture and align images of a structure at different times to identify fatigue cracks through differential image features caused by crack breathing, which can be enhanced and visualized using feature-based and intensity-based image registration, edge-aware noise reduction, and feature enhancement processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If contact-based sensing methods are used for fatigue crack detection, then measurement precision can be improved, but device complexity and ease of operation deteriorate due to extensive human operation required for sensor and actuator deployment
Solution Approach 1:
The patent replaces contact-based mechanical sensing systems with a vision-based optical system. Instead of deploying physical sensors and actuators that require manual installation, the system uses cameras to capture images and processes them through image registration algorithms to detect fatigue cracks, thereby eliminating the complexity of sensor deployment while maintaining detection capability
Solution Approach 2:
The patent creates optical copies (images) of the structure at different times and processes these copies through image registration to detect crack-induced discrepancies. By working with image copies rather than physical sensors, the system avoids the deployment complexity while achieving crack detection through differential analysis
2Measurement precision
If non-destructive testing techniques using acoustic emissions and piezoelectric sensors are used, then measurement precision is improved, but use of energy and device complexity worsen due to additional power requirements and system complexity
Solution Approach 1:
The patent substitutes active sensing systems that require power (acoustic emissions, piezoelectric sensors) with a passive vision-based system. The camera-based approach captures images without requiring additional power for signal generation, eliminating the energy consumption associated with active NDT techniques while maintaining inspection accuracy through image processing
Solution Approach 2:
The system uses ambient light and existing structural features as natural markers for image registration, eliminating the need for external power sources or active illumination. The structure itself provides the reference features needed for alignment, making the system self-sufficient and energy-independent
3Measurement precision
If strain-based monitoring technologies are deployed, then measurement precision improves for detecting fatigue cracks, but device complexity and ease of operation worsen due to extra work required for sensor installation and cabling
Solution Approach 1:
The patent replaces strain-based monitoring that requires physical sensor installation and cabling with a contactless vision-based system. By using cameras to capture and compare images, the system eliminates all installation work associated with mounting strain sensors and running cables, while still achieving crack detection through image differential analysis
Solution Approach 2:
The vision-based system can detect fatigue cracks without requiring any physical contact or installation on the structure, making it universally applicable to various bridge types and locations. The same camera system can inspect multiple structures without reinstallation, providing multi-functionality that strain-based systems lack
4Ease of operation
If human inspection is used to visually examine fatigue cracks, then ease of operation is maintained, but measurement precision and productivity deteriorate due to time consumption, labor intensity, and error proneness
Solution Approach 1:
The patent introduces an intermediary image processing system that bridges simple visual inspection and complex sensor-based detection. The system uses automated image registration and differential analysis to enhance the capabilities of visual inspection, providing machine-assisted precision while maintaining the simplicity of optical observation
Solution Approach 2:
The system creates detailed digital copies of the structure through high-resolution imaging, allowing for precise analysis without requiring physical measurement tools or complex sensor arrays. These image copies enable automated detection algorithms to identify cracks with high precision while keeping the operational interface simple
Data Source
AI summary
An approach for fatigue crack detection is described. In one example, a first image of a structure is captured at a first time, and a second image of the structure is captured at a second time. A feature-based image registration is performed to align features of the second image with the first image, and an intensity-based image registration is performed to further align features of the second image with the first image. A registration error map is determined by performing a pixel-by-pixel intensity comparison of the first image and the second image. Additionally, an edge-aware noise reduction process can be performed on the registration error map. The registration error map can be referenced to identify certain fatigue cracks in the structure.